催化学报  2014, Vol. 35 Issue (8): 1289-1298   PDF (944 KB)    
扩展功能
加入收藏夹
复制引文信息
加入引用管理器
Email Alert
RSS
本文作者相关文章
李倩
谷华春
李萍
周钰浩
刘莹
齐中囡
辛颖
张昭良
In situ IR studies of selective catalytic reduction of NO with NH3 on Ce-Ti amorphous oxides
Qian Lia,b, Huachun Gua, Ping Lia, Yuhao Zhoua, Ying Liua, Zhongnan Qia, Ying Xina, Zhaoliang Zhanga,b     
a. School of Chemistry and Chemical Engineering, University of Jinan, Jinan 250022, Shandong, China;
b. Jiangsu Key Laboratory of Atmospheric Environment Monitoring and Pollution Control, School of Environmental Sciences and Engineering, Nanjing University of Information Sciences and Engineering, Nanjing 210044, Jiangsu, China
Received 31 March 2014. Accepted 23 May 2014. Published 20 August 2014.
* Corresponding author. Tel/Fax: + 86-531-89736032; E-mail: chm_zhangzl@ujn.edu.cn
Foundation Item:This work was supported by the National Natural Science Foundation of China (21277060 and 21107030), the Development Program of the Science and Technology of Shandong Province (2011GSF11702), the Foundation for Outstanding Young Scientist in Shandong Province (BS2011HZ002), the Natural Science Foundation of Shandong Province (ZR2013BQ003), and the Open Project from Jiangsu Key Laboratory of Atmospheric Environment Monitoring and Pollution Control of Nanjing University of Information Science and Technology (KHK1116), Jiangsu Province Innovation Platform for Superiority Subject of Environmental Science and Engineering.
Abstract: A series of in situ infrared (IR) studies of the selective catalytic reduction (SCR) of NOx with NH3 on the short-range ordered structure Ce-O-Ti sites in amorphous Ce-Ti mixed oxides were performed. Under the reaction conditions, the catalyst surface was mainly covered by NH3 ad-species and no NOx ad-species were detected. The reaction order of 0.5-0.6 with respect to NO confirmed a hybrid Langmuir-Hinshelwood and Eley-Rideal mechanism. A possible route may involve the reaction of NH3 ad-species and weakly adsorbed NOx to form an active intermediate, NHyNO3 (y = 0-4); this was confirmed by GAUSSIAN calculations and the in situ IR results. The Ce-O-Ti structure, with Ce-Ti interactions on the atomic scale, enhanced the redox properties in the active temperature window of the SCR reactions.
Key words: Selective catalytic reduction     Nitrogen oxide     Ammonia     Amorphous Ce-Ti mixed oxide     In situ infrared     Mechanism    

1. Introduction

Selective catalytic reduction (SCR) of nitrogen oxides (NOx) with NH3 is not only an important heterogeneous catalytic reaction, but also has immediate practical significance for the control of NOx, an atmospheric pollutant that contributes to acid rain, photochemical smog, and the depletion of tropospheric ozone [1]. The commercial V2O5-WO3(MoO3)/TiO2 (VWTiO2) catalyst causes problems, because of the toxicity of vanadium, high activity in the oxidation of SO2 to SO3, and a relatively narrow temperature window of 300-400 °C [2]. Continuing efforts are therefore being made to develop new catalytic systems.

A comprehensive overview of the NH3-SCR reaction mechanism shows that both surface acidity and redox properties are necessary [1, 3]. This strategy has been used to develop many new catalysts [4, 5, 6, 7, 8, 9]. Recently, Ce-based catalysts, particularly Ce-Ti mixed oxides, have been receiving much attention [10, 11, 12, 13, 14, 15, 16, 17, 18]. However, little has been reported on the reaction mechanism and kinetics.

The SCR reaction mechanism is often studied using in situ infrared (IR) spectroscopy. Regardless of the catalyst used, the rate constant for evaluating the activity is often assigned based on the hypothesis that the SCR reaction is approximately first order with respect to NO [8, 19]. Kinetic data need to be obtained to determine the validity of this assumption.

In VWTiO2 systems, the active acid sites (V-OH) and redox sites (V=O) are separate on the catalyst surface [3]. However, if they are combined, synergism provides more efficient catalysis. Previously, we proposed that the short-range ordered structures in the amorphous phase could be good candidates for this. In our previous work, this was achieved by forming Ce-O-Ti species with Ce-Ti interactions on the atomic scale [20]. In the present work, further mechanistic aspects are investigated using in situ IR absorption/reaction spectra and kinetic data.

2. Experimental
2.1. Catalyst preparation

A series of Ce-Ti mixed oxides with different molar ratios of Ce and Ti were prepared using a co-precipitation method [20]; these are denoted by CeaTiOx, where a represents the Ce/Ti atomic ratio and equals 0.3, 0.5, or 0.7. For comparison, CeO2 and TiO2 were also prepared.

2.2. Catalyst test

A fixed-bed U-shaped quartz reactor (ID = 6 mm) with a thermocouple placed inside the catalyst was used for kinetic reaction measurements under atmospheric conditions. A model flue gas consisting of 0.05% NO, 0.05% NH3, and 5.3 vol% O2 in He at 300 mL min-1, and a gas hourly space velocity (GHSV) of 100000 h-1 was used. In order to ensure that only the intrinsic SCR reactions are in the kinetic regime, a series of experiments were performed to exclude internal and external diffusion and heat transfer effects. We found that there were no intraparticle mass transport limitations when the diameter was below 100-200 mesh. For a total flow rate of 300 mL min-1, no external mass transport limitations were detected. When the conversion of NOx was lower than 20%, there was no increase in the temperature of the catalyst bed. The concentrations of NO and NO2 were monitored using a chemiluminescence NOx analyzer (42i-HL, Thermo). N2O and NH3 were detected using a Fourier-transform IR spectrometer (Tensor 27, Bruker) with a gas cell with a 2.4 m optical path.

2.3. Catalyst characterization

In situ X-ray diffraction (XRD) patterns were collected at the XRD station of the Beijing Synchrotron Radiation Facility, using a curved image plate detector with an angle accuracy of 0.01° and an X-ray wavelength of 1.38 Å, in transmission mode. Temperature-programmed reduction with H2 (H2-TPR) experiments were performed in a quartz reactor with a thermal conductivity detector to monitor the H2 consumed. A 50 mg sample was pretreated in situ at 500 °C for 1 h in a flow of O2, and cooled to room temperature in the presence of O2. H2-TPR experiments were conducted at 10 °C min-1 up to 900 °C in a 30 mL min-1 flow of 5 vol% H2 in N2. To quantify the total amount of H2 consumed, CuO was used as a calibration reference. The temperature-programmed desorption of NH3 (NH3-TPD) experiments were performed in a quartz reactor using 50 mg catalyst. NH3 was monitored using quadrupole mass spectrometer (MS; OmniStar 200, Balzers) with m/z = 16. Prior to the experiments, the samples were pretreated at 500 °C for 30 min in 10 vol% O2/He (50 mL min-1) and cooled to 100 °C. NH3 adsorption was operated in 0.4% NH3 (50 mL min-1) until the outlet NH3 concentration was stable. Then, the samples were purged with He to remove any weakly absorbed NH3. Finally, the samples were heated to 700 °C at a ramping rate of 10 °C min-1.

In situ IR spectra were recorded using a Bruker Tensor 27 spectrometer over the range 400-4000 cm-1, with 16 scans, at a resolution of 4 cm-1. Self-supporting wafers were pretreated in the IR cell at 500 °C in a flow of He for 30 min to remove any adsorbed species. After cooling to 100 °C, the background spectrum was recorded. Adsorption/reaction spectra were then recorded at the desired temperatures, as indicated in the figure captions. Because Ce0.7TiOx was difficult to mold, an approximately double-mass sample, compared to Ce0.3TiOx and Ce0.5TiOx, was pelletized and then used in the in situ IR experiments. In some cases, the effluents from the in situ IR cell were also monitored, using a quadrupole mass spectrometer (MS; OmniStar 200, Balzers) with m/z values of 28 for N2, 17 for NH3, 30 for NO, and 46 for NO2.

2.4. Theoretical calculation

The GAUSSIAN 09 program was used to optimize the structures and calculate the vibrational wavenumbers for the NxHyOz intermediate on CeaTiOx. The 6-311++G** basis set was used to perform the DFT-B3PW91 calculations (Becke’s three- parameter functional with the non-local correlation provided by the Perdew 91 expression).

3. Results and discussion

Our previous work [20] and in situ synchrotron-radi­ation XRD patterns (Fig. 1) show an amorphous structure and the presence of possible short-range ordered structures of Ce-O-Ti and Ti-O-Ti in Ce0.3TiOx, Ce-O-Ti in Ce0.5TiOx, and Ce-O-Ti and Ce-O-Ce in Ce0.7TiOx calcined at 500 °C, in accordance with the fact that the Ce/Ti atomic ratios are less than, equal to, and more than the stoichiometry of CeTi2O6.

Fig. 1. In situ heating XRD patterns for CeaTiOx. (a) Ce0.3TiOx; (b) Ce0.5TiOx; (c) Ce0.7TiOx.

Fig. 2 shows NH3-TPD curves on TiO2, CeO2, and CeaTiOx adsorbed at 100 °C after purging with He, and the corresponding IR spectra are shown in Fig. 3. NH3 desorption on CeO2 is negligible. However, TiO2 and CeaTiOx show broad NH3 desorption peaks between about 100 and 500 °C. NH3 adsorption species on Brönsted acid sites derived from residual sulfate from the precursor Ti(SO4)2 were not observed [18, 21]. TiO2 has only surface Lewis acid sites[22], and no NH4+ species can be formed on CeO2 [23], therefore the common bands at 1440 and 1670 cm-1 for CeaTiOx can be assigned to NH4+ on Ce-O-Ti Brönsted acid sites, and the bands at 1310, 1292, and 1297 cm-1 arise from coordinated NH3 on Ce-O-Ti Lewis acid sites (Table 1). The bands at 1600 and 1176 cm-1 were only observed for Ce0.3TiOx and TiO2, and are assigned to coordinated NH3 on Ti-O-Ti Lewis acid sites. The present results indicate that there were more Brönsted acid sites than Lewis acid sites on the Ce-O-Ti structure at 100 °C. With increasing temperature, the band at 1440 cm-1 gradually weakened and eventually disappeared at 250, 350, and 400 °C for Ce0.3TiOx, Ce0.5TiOx, and Ce0.7TiOx, respectively. Simultaneously, the bands for Lewis acid sites shifted to higher wavenumbers, eventually reaching 1350 cm-1 at 450 °C for all CeaTiOx samples.

Fig. 2. NH3-TPD curves of TiO2, CeO2, and CeaTiOx.

Fig. 3. In situ IR spectra of NH3 adsorbed on CeaTiOx as a function of temperature.

Table 1
IR bands and assignments.

Figure 4 shows a series of IR spectra of NO + O2 adsorbed on CeaTiOx at different temperatures. At 100 °C, strong IR bands were observed at 1617, 1555, and either 1235, 1285, or 1297 cm-1. The first band is attributed to an NO2 ad-species (nitro group or adsorbed NO2 molecule) on Ce-O-Ti sites, because it did not disappear after purging with He for 30 min (Fig. 5), the second corresponds to bidentate nitrate, and the last three correspond to bridge-bound nitrites/nitrates on Ce-O-Ti sites. The band at 1576 cm-1 was only observed for Ce0.3TiOx and is assigned to bidentate nitrate on Ti-O-Ti sites. With increasing temperature, these bands vanished and a new band at about 1360 cm-1 was formed at 450 °C for all CeaTiOx samples.

Fig. 4. In situ IR spectra of NO + O2 adsorbed on CeaTiOx as a function of temperature.

Fig. 5. In situ IR spectra of NO + O2 adsorption on CeaTiOx at 100 °C for 30 min, followed by purging with He for 30 min.

Figure 6 shows IR spectra of the reaction between NO + O2 and NH3 ad-species at 100 °C. For Ce0.3TiOx, both NH4+ and coordinated NH3 species disappeared from the surface after NO + O2 was passed over the sample for 15 min (Fig. 6(a)). In addition, the disappearance of coordinated NH3 on Ti-O-Ti Lewis acid sites (1176 and 1600 cm-1) indicates that the Ti-O-Ti structure (TiO2) is capable of carrying out the reduction of NO by NH3; however, this is not a catalytic reaction, because TiO2 does not show any catalytic activity at this temperature [20, 30]. For Ce0.5TiOx, the reaction between NO + O2 and NH3 ad-species requires about 45 min (Fig. 6(b)). However, this reaction was incomplete even after 70 min for Ce0.7TiOx. These results suggest that (1) NH4+ ions and coordinated NH3 on the Ce-O-Ti structure are both active in SCR reactions and (2) the activity order is Ce0.3TiOx > Ce0.5TiOx > Ce0.7TiOx.

Fig. 6. In situ IR spectra of surface reactions on CeaTiOx between pre-adsorbed NH3 and NO + O2 at 100 °C.

Figure 7 shows IR spectra of the reaction between NH3 and NO + O2 ad-species at 100 °C. After NH3 introduction into the cell for 2-3 min, the intensities of the IR bands from NO2 ad-species at 1617 cm-1 and nitrate species quickly decreased, corresponding to the formation of N2, as observed by on-line MS. This indicated that NOx ad-species had reacted with NH3. IR bands attributed to NH4+ ions and coordinated NH3 on the Ce-O-Ti structure were formed.

Fig. 7. In situ IR spectra of surface reactions on CeaTiOx between pre-adsorbed NO + O2 and NH3.

Figure 8 shows IR spectra of CeaTiOx in a flow of NO + NH3 + O2. At 100 °C, the surface was covered mainly by NH4+ ions in comparison with those in Fig. 5. Raising the temperature resulted in a decrease in the intensity of the peak from NH4+ ions, which finally vanished at 250, 350, and 400 °C for Ce0.3TiOx, Ce0.5TiOx, and Ce0.7TiOx, respectively. For coordinated NH3 species, the IR bands shifted gradually to about 1350 cm-1 at 450 °C, similar to those obtained in Fig. 3 and those obtained in the adsorption of NO + O2, which are always in the range 1355-1361 cm-1 from a low temperature to 450 °C (Fig. 4). This suggests that the coordinated NH3 species are similar to the monodentate nitrite species or nitro compounds on Ce-O-Ti sites, e.g., NxHyOz [26]. IR bands from NOx ad-species were not detected across the entire temperature range under the reaction conditions used, except for two barely detectable bands at 1906 and 1850 cm-1 ascribed to gaseous or weakly adsorbed NOx; this is typical behavior for SCR reactions that proceed via the Eley-Rideal (ER) mechanism [21, 24]. This suggests that adsorption of NH3 was dominant under the SCR reaction conditions, as shown in Fig. 7.

Fig. 8. In situ IR spectra of NO + NH3 + O2 reactions on CeaTiOx. 0.05% NO + 0.05% NH3 + 5.3 vol% O2 + He, total flow 120 mL min-1.

As shown in our previous work, the order of catalytic activity in NOx conversion is Ce0.3TiOx > Ce0.5TiOx > Ce0.7TiOx [20], which is the same as that for the decrease in Tmax (the peak temperature for H2 consumption) in H2-TPR (Tmax and the amount of H2 consumed for Ce0.7TiOx are 710 °C and 906 μmol g-1, respectively). This confirms the indispensable role of the catalyst redox properties, which catalyze NO oxidation to NOx ad-species. As can be seen from Fig. 4, the oxidation activity follows the same order as the NOx conversion and H2-TPR properties. This has been confirmed to be derived from the formation of Ce-O-Ti short-range ordered structures with Ce-Ti interaction on the atomic scale. The reducibility plays a key role in NH3 activation and NO oxidation [31, 32].

Steady-state kinetic studies were performed to determine the reaction mechanism. A large excess of O2 was used, therefore the reaction order with respect to O2 is assumed to be 1. The strong adsorption of NH3 (Fig. 8) indicates a zero-order reaction with respect to NH3. The kinetic reaction rate of NOx conversion as a function of reactant concentrations can be expressed as rNO conv. = k[NO]x. The reaction orders with respect to NO were calculated to be 0.67, 0.58, and 0.67 for Ce0.3TiOx, Ce0.5TiOx, and Ce0.7TiOx, respectively (Fig. 9). All of these values are less than 1, suggesting that the SCR reaction involves both the Langmuir-Hinshelwood (LH) and ER pathways [33]. This is commonly observed for low-temperature SCR oxide catalysts [19]. However, as shown in Fig. 8, no NOx ad-species were detected. The IR spectra were very similar to the NH3 adsorption curves, even though NOx adsorption did take place in the NO + O2 flow without NH3 (Fig. 4). This strongly suggests that the activity of the reaction between NH3 and NOx ad-species is much higher than that between NO + O2 and NH3 ad-species; this is confirmed by the IR spectra of the two reactions (Fig. 6 vs Fig. 7). The pathway possibly involves the reaction of NH3 ad-species and weakly adsorbed NOx. A similar mechanism on CeO2/TiO2 and its W-doped analog has been observed [26].

Fig. 9. Reaction orders with respect to NO for CeaTiOx. Reaction conditions: 0.02-0.1% NO + 0.1% NH3 + 5.3 vol% O2 + He, total flow 300 mL min-1, GHSV 100000 h-1, reaction temperature 160 °C.

Possible structures of NxHyOz were determined using the GAUSSIAN program [34]. As shown in Table 2, the calculated NO stretching frequency is 1359 cm-1, which is close to the experimental data in Fig. 3 (1352, 1353, and 1346 cm-1), Fig. 4 (1356, 1361, and 1357 cm-1), and Fig. 8 (1352, 1353, and 1350 cm-1). A specific compound, NH4NO3, has been identified as an intermediate by many researchers [35]. Furthermore, the NHd bending vibration at 1535 cm-1 is close to those observed in Fig. 6 (1522 cm-1), Fig. 7 (1520 cm-1), and Fig. 8 (1535 cm-1). It is deduced that NHyNO3 is one of the most probable intermediates in the present case. As shown in Figs. 3 and 8, below temperatures at which NH4+ species (1440 cm-1) existed (~250, ~350, and ~400 °C for Ce0.3TiOx, Ce0.5TiOx, and Ce0.7TiOx, respectively), y ≤ 4. With increasing temperature, y gradually decreases, even reaching 0, because no NH3 can be desorbed above ~450°C (Fig. 2), i.e., the oxidation of NH3 to nitrites/ nitrate species is predominant (Figs. 3 and 4). This corresponds to a decrease in NOx conversion to N2.

Table 2
Structure and calculated vibrational frequencies and intensities of NxHyOz.
4. Conclusions

On Ce-Ti amorphous oxides, Ce-O-Ti short-range ordered structures with Ce-Ti interactions on the atomic scale enhanced NO oxidation and NH3 activation. Under these reaction conditions, the catalyst surface was mainly covered by NH3 ad-species, and no NOx ad-species were detected. The reaction order with respect to NO was calculated to be 0.5-0.6, supporting a hybrid LH and ER mechanism. A possible route involves the reaction of NH3 ad-species and weakly adsorbed NOx to form an active intermediate, NHyNO3 (y = 0-4); this was confirmed by GAUSSIAN calculations and in situ IR data.

References
[1] Busca G, Lietti L, Ramis G, Berti F. Appl Catal B, 1998, 18: 1
[2] Kompio P G W A, Brückner A, Hipler F, Auer G, Löffler E, Grünert W. J Catal, 2012, 286: 237
[3] Topsøe N Y. Science, 1994, 265: 1217
[4] Li F, Zhang Y B, Xiao D H, Wang D Q, Pan X Q, Yang X G. ChemCat-Chem, 2010, 2: 1416
[5] Zhang Y B, Wang D Q, Wang J, Chen Q F, Zhang Z D, Pan X Q, Miao Z Z, Zhang B, Wu Z J, Yang X G. Chin J Catal (张一波, 王德强, 王静, 陈去非, 张震东, 潘喜强, 苗珍珍, 张彬, 武志坚, 杨向光. 催化学报), 2012, 33: 1448
[6] Liu F D, He H, Ding Y, Zhang C B. Appl Catal B, 2009, 93: 194
[7] Shan W P, Liu F D, He H, Shi X Y, Zhang C B. Chem Commun, 2011, 47: 8046
[8] Huang Z W, Gu X, Wen W, Hu P P, Makkee M, Lin H, Kapteijn F, Tang X F. Angew Chem Int Ed, 2013, 52: 660
[9] Mou X L, Zhang B S, Li Y, Yao L D, Wei X J, Su D S, Shen W J. Angew Chem Int Ed, 2012, 51: 2989
[10] Xu W Q, Yu Y B, Zhang C B, He H. Catal Commun, 2008, 9: 1453
[11] Gao X, Jiang Y, Zhong Y, Luo Z Y, Cen K F. J Hazard Mater, 2010, 174: 734
[12] Gao X, Jiang Y, Fu Y C, Zhong Y, Luo Z Y, Cen K F. Catal Commun, 2010, 11: 465
[13] Shan W P, Liu F D, He H, Shi X Y, Zhang C B. ChemCatChem, 2011, 3: 1286
[14] Gao X, Du X S, Cui L W, Fu Y C, Luo Z Y, Cen K F. Catal Commun, 2010, 12: 255
[15] Jin R B, Liu Y, Wu Z B, Wang H Q, Gu T T. Catal Today, 2010, 153: 84
[16] Chen L, Li J H, Ge M F, Zhu R H. Catal Today, 2010, 153: 77
[17] Wu D W, Zhang Q L, Lin T, Gong M C, Chen Y Q. Chin J Inorg Chem (吴大旺, 张秋林, 林涛, 龚茂初, 陈耀强. 无机化学学报), 2011, 27: 53
[18] Xu W Q, He H, Yu Y B. J Phys Chem C, 2009, 113: 4426
[19] Yang S J, Wang C Z, Li J H, Yan N Q, Ma L, Chang H Z. Appl Catal B, 2011, 110: 71
[20] Li P, Xin Y, Li Q, Wang Z P, Zhang Z L, Zheng L R. Environ Sci Tech-nol, 2012, 46: 9600
[21] Yang S J, Guo Y F, Chang H Z, Ma L, Peng Y, Qu Z, Yan N Q, Wang C Z, Li, J H. Appl Catal B, 2013, 136-137: 19
[22] Wachs I E, Deo G, Weckhuysen B M, Andreini A, Vuurman M A, de Boer M, Amiridis M D. J Catal, 1996, 161: 211
[23] Qi G S, Yang R T, Chang R. Appl Catal B, 2004, 51: 93
[24] Long R Q, Yang R T. J Catal, 2000, 190: 22
[25] Fornasari G, Trifirò F, Vaccari A, Prinetto F, Ghiotti G, Centi G. Catal Today, 2002, 75: 421
[26] Chen L, Li J H, Ge M F. Environ Sci Technol, 2010, 44: 9590
[27] Yung M M, Holmgreen E M, Qzkan U S. J Catal, 2007, 247: 356
[28] Wu Z B, Jiang B Q, Liu Y, Wang H Q, Jin R B. Environ Sci Technol, 2007, 41: 5812
[29] Qi G S, Yang R T. J Phys Chem B, 2004, 108: 15738
[30] Topsøe N Y, Dumesic J A, Topsøe H. J Catal, 1995, 151: 241
[31] Liu Z M, Yi Y, Li J H, Woo S I, Wang B Y, Cao X H, Li Z X. Chem Commun, 2013, 49: 7726
[32] Peng Y, Li J H, Huang X, Li X, Su W K, Sun X X, Wang D Z, Hao J M. Environ Sci Technol, 2014, 48: 4515
[33] Sjoerd Kijlstra W, Brands D S, Poels E K, Bliek A. Catal Today, 1999, 50: 133
[34] Kebede M A, Varner M E, Scharko N K, Gerber R B, Raff J D. J Am Chem Soc, 2013, 135: 8606
[35] Li F D, He H. Catal Today, 2010, 153: 70
非晶Ce-Ti氧化物用于NH3选择性催化还原NO的原位红外研究
李倩a,b, 谷华春a, 李萍a, 周钰浩a, 刘莹a, 齐中囡a, 辛颖a, 张昭良a,b     
a. 济南大学化学化工学院, 山东济南 250022;
b. 南京信息工程大学环境科学与工程学院, 江苏省大气环境监测与污染控制高技术研究重点实验室, 江苏南京 210044
收稿日期: 2014-3-31. 接受日期: 2014-5-23. 出版日期: 2014-8-20.
基金来源：国家自然科学基金（21277060和21107030）；山东省科技发展计划项目（2011GSF11702）；山东省优秀中青年科学家科研奖励基金（BS2011HZ002）；山东省自然科学基金（ZR2013BQ003）；江苏省大气环境监测与污染控制高技术研究重点实验室开放基金（KHK1116），江苏省环境科学与工程优势学科平台.
*通讯联系人.张昭良, 电话: (0531) 89736032; 传真: (0531) 89736032; 电子信箱: chm_zhangzl@ujn.edu.cn
摘要：采用原位红外光谱研究了在具有短程有序Ce-O-Ti结构的非晶Ce-Ti氧化物上NH3选择性催化还原（SCR） NOx反应. 在反应条件下，催化剂表面主要被NH3吸附物种覆盖，而检测不到NOx吸附物种. 经测定，NO的反应级数为0.5-0.6，表明Langmuir-Hinshelwood机理和Eley-Rideal机理同时存在. 可能的机理是NH3吸附物种和弱吸附的NOx反应，生成NHyNO3 （y = 0-4）活性中间物种，并通过GAUSSIAN计算和原位红外结果证实了它们的存在. Ce-O-Ti结构中Ce与Ti之间表现出原子尺度的相互作用，所以在SCR反应的活性温度窗口下，催化剂的氧化还原活性提高.
关键词： 选择性催化还原     一氧化氮     氨气     非晶Ce-Ti氧化物     原位红外     机理    
1. 前言

氮氧化物(NOx)的危害很大, 可产生酸雨、光化学烟雾和臭氧层空洞等, 氨气选择性催化还原(NH3-SCR)技术是有效控制NOx的重要手段[1].  商用V2O5- WO3(MoO3)/TiO2(VWTiO2)催化剂中的钒具有毒性, 其催化氧化SO2为SO3的活性过高, 活性温度窗口也比较窄(300-400 °C)[2].  因此, 需要努力开发新型催化剂体系.  

NH3-SCR反应机理研究表明催化剂的表面酸性和氧化还原性对反应性能影响很大[1, 3].  据此开发了许多新型催化剂[4, 5, 6, 7, 8, 9].  最近, Ce基催化剂, 尤其是Ce-Ti复合氧化物受到极大关注[10, 11, 12, 13, 14, 15, 16, 17, 18].  但对反应机理和动力学的报道相对较少.  

文献中通常采用原位红外(in situ IR)技术来研究SCR的反应机理.  但是, 不论选用何种催化剂, 用于评价活性的速率常数都是基于SCR反应对NO的级数接近于1的假设[8, 19].  因此, 需要提供动力学数据来证实这种假设.  

对于VWTiO2体系, 催化剂上的活性酸性位(V-OH)和氧化还原位(V=O)是独立的[3].  但是如果能将它们统一到一起, 协同效应将会更加显著.  我们曾通过形成Ce-O-Ti物种, 使得Ce与Ti之间存在原子尺度的相互作用, 从而在非晶相中形成短程有序结构, 可导致上述效应增强[20].  本文采用in situ IR吸附/反应光谱和动力学数据进一步对反应机理进行了研究.

2. 实验部分
2.1. 催化剂制备

采用共沉淀法制备了一系列不同Ce/Ti摩尔比的铈钛复合氧化物[20], 将它们标记为CeaTiOx, 其中a表示Ce/Ti原子比, 分别为0.3, 0.5和0.7.  为了比较, 同法制备了CeO2和TiO2.  

2.2. 催化剂测试

动力学反应测试在常压固定床U形石英管反应器上进行.  催化剂置于内径为6 mm的反应管中, 并插入热电偶测量催化剂温度.  模拟烟气总流量为300 mL min-1, 其中含有0.05% NO, 0.05% NH3, 5.3 vol% O2, He为平衡气, 体积空速(GHSV)为100000 h-1.  为了确保SCR本征反应在动力学模式下进行, 我们进行了一系列的实验以排除内扩散、外扩散和传热因素的影响.  发现当催化剂粒径低于100-200目时, 颗粒内部质量传递不受限制.  反应气流量达到300 mL min-1时, 外部的质量传递不受限制.  当NOx的转化率低于20%时, 催化剂床层温度没有明显的升高.  NO和NO2的浓度通过化学发光NOx分析仪(42i-HL, Thermo)检测.  N2O和NH3通过傅里叶变换红外(FTIR)光谱仪(Tensor 27, Bruker)进行检测, 其中气体池的光路长度为2.4 m.  

2.3. 催化剂表征

原位XRD测试在北京同步辐射装置的XRD站上进行, 采用弯曲的图像板检测器, 测角精度为0.01°, X射线波长为1.38 Å, 测试在透射模式下进行.  氢气程序升温还原(H2-TPR)实验在石英管反应器中进行, TCD检测氢气消耗.  样品用量为50 mg, 在O2气氛下500 oC原位预处理1 h, 然后在O2气氛下降至室温.  然后切换为5%H2/N2混合气(30 mL min-1), 以10 °C min-1升至900 oC.  采用CuO为标准样对样品的耗氢量进行定量计算.  氨气程序升温脱附实验(NH3-TPD)在石英管反应器中进行, 催化剂用量为50 mg.  利用瑞士BALZERS公司生产的Omnistar质谱仪检测NH3的浓度, 核质比采用16.  实验前, 将样品在10 vol% O2/He (50 mL min-1)中500 °C预处理30 min.  预处理结束后, 降温至100 °C, 通入0.4%的NH3 (50 mL min-1)进行吸附.  待吸附饱和后, 通入高纯氦气吹扫至质谱信号稳定.  最后, 以10 °C min-1升至700 °C进行脱附.  

In situ IR实验在美国Bruker公司Tensor 27红外光谱仪上进行, 扫谱范围为400-4000 cm-1, 扫描次数为16次, 分辨率为4 cm-1.  自支撑片在红外池中500 °C下He气氛下预处理30 min以除去吸附物种.  然后降至100 °C并录背景谱, 在给定温度下记录吸附/反应光谱.  由于Ce0.7TiOx难以压片, 因而其用量约为Ce0.3TiOx或Ce0.5TiOx的两倍.  某些情况下, 原位红外池中排出的气体也通过四极质谱(MS, OmniStar 200, Balzers)进行检测, 其中N2, NH3, NO和NO2的m/z分别为28, 17, 30和46.  

2.4. 理论计算

使用GAUSSIAN 09程序优化CeaTiOx中NxHyOz中间物种的结构, 并计算其振动波数.  采用密度泛函理论B3PW91方法(Becke三参数混合泛函, 非局域相关泛函由Perdew 91提供), 6-311++G**基组进行计算.  

3. 结果与讨论

前期工作[20]和原位同步辐射XRD (图1)结果表明:  500 °C焙烧时, 催化剂为非晶结构, 其中可能含有短程有序的结构, Ce0.3TiOx中为Ce-O-Ti和Ti-O-Ti, Ce0.5TiOx中为Ce-O-Ti, Ce0.7TiOx中为Ce-O-Ti和Ce-O-Ce.  这与样品中的Ce/Ti原子比是一致的, 三种样品的Ce/Ti原子比分别小于、等于和大于CeTi2O6的化学计量比.  

图2为TiO2, CeO2和CeaTiOx上100 °C吸附NH3并用He吹扫后的程序升温脱附(TPD)谱图, 相应的随温度变化的NH3的IR谱见图3.  可以看出, CeO2上几无NH3的脱附, 而TiO2和CeaTiOx在100-500 °C范围内出现一个宽的NH3脱附峰.  实验中并没有观察到可能由于制备过程中Ti(SO4)2前驱体残留的硫酸盐而导致的Brönsted酸性位上的NH3吸附物种[18, 21].  由于TiO2仅含有表面Lewis酸性位[22], CeO2上也没有生成NH4+物种[23], 所以可将CeaTiOx上的1440和1670 cm-1的峰归属于Ce-O-Ti中Brönsted酸性位上的NH4+物种, 而1310, 1292和1297 cm-1的峰归属为Ce-O-Ti中Lewis酸性位上的配位态NH3物种(表1).  仅在Ce0.3TiOx和TiO2样品上观测到了1600和1176 cm-1的峰, 将其归属为Ti-O-Ti中Lewis酸性位上的配位态NH3物种.  结果表明:  在Ce-O-Ti结构中, 100 °C时Brönsted酸性位比Lewis酸性位更多.  Ce0.3TiOx, Ce0.5TiOx和Ce0.7TiOx样品上, 随着温度的升高, 1440 cm-1处的峰逐渐弱化, 并分别于250, 350和400 °C时最终消失.  同时对于所有的CeaTiOx样品, Lewis酸性位上的峰逐渐向高波数移动, 450 °C时最终移动至1350 cm-1处.  

图4给出了CeaTiOx在不同温度下吸附NO + O2的IR谱.  100 °C时, 观测到了1617, 1555和1235/1285/1297 cm-1处的强峰.  第一个峰在He气氛下吹扫30 min后仍不消失(图5), 将其归属于Ce-O-Ti位上的NO2吸附物种(硝基或吸附态NO2分子), 第二个峰对应着双齿硝酸盐, 最后三个峰归属于Ce-O-Ti位上的桥式键合的亚硝酸盐或硝酸盐物种.  在Ce0.3TiOx上仅观测到了1576 cm-1处的峰, 将其归属于Ti-O-Ti位上的双齿硝酸盐.  对于所有的CeaTiOx样品, 随着温度的升高, 这些峰都逐渐消失, 并在450 °C时于1360 cm-1处出现了一个新峰.  

图6为NO + O2与NH3吸附物种在100 °C下反应的IR谱.  对于Ce0.3TiOx, NO + O2通入后15 min, 表面上的NH4+物种和配位态的NH3物种均消失(图6(a)).  另外, Ti-O-Ti中Lewis酸性位上配位态NH3物种(1176和1600 cm-1)消失, 表明Ti-O-Ti结构(TiO2)上NH3可以还原NO.  但是, TiO2在此温度下没有催化活性[20, 30], 所以认为该过程并不是催化过程.  对于Ce0.5TiOx, NO+ O2与NH3&l t;/ span>吸附物种的反应需要大约45 min (图6(b)).  但是对于Ce0.7TiOx, 反应进行70 min后仍未结束.  这些结果表明:  (1)在SCR反应过程中, Ce-O-Ti结构上的NH4+离子和配位态的NH3均具有活性;  (2)活性顺序为Ce0.3TiOx > Ce0.5TiOx > Ce0.7TiOx.  

图7为100 °C下NH3与NO + O2吸附物种反应的IR谱.  NH3通入红外池2-3 min后, NO2吸附物种(1617 cm-1)和硝酸盐物种的红外峰迅速减弱, 同时对应着N2的生成(通过在线质谱检测).  这表明NOx吸附物种与NH3进行了反应.  同时, Ce-O-Ti结构上出现了属于NH4+离子和配位态NH3的红外峰.  

图8为CeaTiOx上NO + NH3 + O2气氛下的IR谱.  100 °C时, 与图5相比, 发现CeaTiOx表面主要被NH4+离子覆盖;  提高温度导致该离子峰强度降低, 在Ce0.3TiOx, Ce0.5TiOx和Ce0.7TiOx上, 此峰分别于250, 350和400 °C时最终消失.  对于配位态NH3物种, 红外峰逐渐移至约1350 cm-1处(450 °C), 与NH3-TPD结果类似(图3), 也与图4的&l t;/ span>NO + O2吸附实验中总会在1355-1361 cm-1范围内出现红外峰的结果类似.  这表明配位态的NH3物种与Ce-O-Ti位上的单齿亚硝酸盐物种或硝基化合物(如NxHyOz)[26]具有一定的相似性.  在反应过程中的整个温度区间内, 除了在1906和1850 cm-1处有两个不可分辨的IR峰以外(可以归属为气相或弱吸附的NOx物种), 并没有观测到NOx吸附物种的红外峰, 这是通过ER机制进行的SCR反应的典型特点[21, 24].  由此可见在SCR反应条件下, NH3的吸附是最主要的(图7).  

前期研究发现, 各催化剂上NOx转化率的顺序为:  Ce0.3TiOx > Ce0.5TiOx > Ce0.7TiOx[20], 这与H2-TPR实验中Tmax (耗氢峰顶温度)降低的顺序是一致的(Ce0.7TiOx样品的Tmax与耗氢量分别为710 oC和906 μmol g-1).  这表明催化剂的氧化还原特性起着不可或缺的作用, 可能表现为NO氧化为NOx吸附物种的能力.  如图4所示, 氧化活性同NOx转化率和H2-TPR特性的顺序一致.  这可由Ce-O-Ti短程有序结构的生成(Ce与Ti存在原子 尺度的相互作用)得到证实.  催化剂的可还原性在NH3活化和NO氧化过程中起着关键的作用[31, 32].  

为了确认反应机理, 我们进行了稳态动力学研究.  由于反应过程中O2是大大过量的, 所以反应对O2的级数可以认为是1.  同时, 由于NH3存在强吸附(图8), 其反应级数可以认为是0.  NOx转化的动力学反应速率与反应物浓度的函数关系可以表示为:  rNO conv. = k [NO]x.  通过计算, Ce0.3TiOx, Ce0.5TiOx和Ce0.7TiOx上NO的反应级数分别为0.67, 0.58和0.67 (图9).  所有的数值均小于1, 表明SCR反应同时包含LH和ER路径[33].  这是低温SCR氧化物催化剂常见的特性[19].  但是, 图8中没有检测到NOx吸附物种.  即便在无NH3的NO + O2气氛中(图4)催化剂上的确发生了NOx的吸附, 但图8中的红外光谱与NH3的吸附光谱是非常相似的.  这充分表明, NH3与NOx吸附物种的反应活性远远高于NO + O2</ sub>与NH3吸附物种的, 这也可以从两种反应的IR谱得到证实(图6和图7).  因此, 反应路径可能包含NH3吸附物种与弱吸附NOx的反应.  在CeO2/TiO2和W掺杂的催化剂上也证实存在类似的机理[26].  

使用GAUSSIAN程序计算了NxHyOz的可能结构[34].  如表2所示, 算得NO伸缩振动频率为1359 cm-1, 这与图3 (1352, 1353和1346 cm-1)、图4 (1356, 1361和1357 cm-1)和图8(1352, 1353和1350 cm-1)中的实验数据非常接近.  许多研究者认为, NH4NO3是反应的中间物种[35].  另外, 1535 cm-1处NHd弯曲振动峰与图6 (1522 cm-1)、图7 (1520 cm-1)和图8 (1535 cm-1)中的位置是很接近的.  & #8197;因此可以推断, NHyNO3是本文反应中可能性最大的中间物种之一.  如图3和图8所示, 低于NH4+物种(1440 cm-1)存在的温度时(Ce0.3TiOx, Ce0.5TiOx和Ce0.7TiOx分别为~250, ~350和~400 °C), y等于4或者更小.  随着温度的升高 , y逐渐降低甚至降为0 (如图2所示高于~450 °C时没有NH3脱附).  也就是说, 这种情况下NH3氧化为亚硝酸盐/硝酸盐物种是占主导作用的(图3和图4);  相应地, NOx生成N2的转化率下降.  

4. 结论

在Ce-Ti非晶氧化物上, Ce-O-Ti短程有序结构中Ce与Ti存在原子尺度的相互作用, 可以促进NO的氧化和NH3的活化.  反应条件下, 催化剂表面主要被NH3吸附物种覆盖而没有检测到NOx吸附物种.  经计算得到反应对NO的级数为0.5-0.6, 表明Langmuir- Hinshelwood和Eley-Rideal机理共存.Q ; 97; 反应路径可能包含NH3吸附物种与弱吸附的NOx反应生成NHyNO3 (y = 0-4)活性中间物种, 此推论可为GAU­SSIAN计算和原位红外结果所证实.